The Composition of Fusel Oil from Beet Molasses. - Industrial

Cesar A. Sánchez , Orlando A. Sánchez , Alvaro Orjuela , Iván D. Gil , and Gerardo Rodríguez. Journal of Chemical & Engineering Data 2017 62 (1), 11-1...
1 downloads 0 Views 417KB Size
Sept., 1 9 1 2

T I I E JOCR.YAL OF I.\7DC.STRIAL

silicate of aluminum. while the latter contains relatively more silica and iron and correspondingly less aluminum. I n some instances i t was found t h a t a coagulation could be affected b y the use of larger amounts of ammonium carbonate, while in other cases, as much as 40 grams per liter failed t o produce coagulation. By the use of still larger amounts of ammonium carbonate, a partial precipitation of humus also took place. COXTRIBDTION h-0. 5 , HAWAIIEXPERIMENT STATION, HONOLULU.

THE COMPOSITION OF FUSEL OIL FROM BEET MOLASSES. B y M. c. BOSWELLAND J. L. GOODERHAM?. Received May 1 5 . 1912.

The published proximate analyses of fusel oil are very few and mostly relate t o the distillates from potatoes and corn. This arises from the fact t h a t , until comparatively recent years, fusel oil from any source was regarded as more or less valueless. I n Canada, after legislation put a premium on its destruction, this product was allowed t o run away in drains or used as a constituent of a cheap burning fluid. The numerous purposes t o which it is now applied and for which i t is indeed indispensable, have changed this custom, so t h a t it has become four or five times as valuable as the alcohol from which it is separated. The production of fusel oil from the fermentation of beet molasses depends industrially on a supply of crude material from beet sugar refineries, which did not come into existence on a large scale in Canada until about seven years ago. An alcohol plant, tons of capable of handling from ~o,ooo-~o,ooo molpses per annum, was established in Tcironto in 1 9 0 j , and very considerable quantities of fusel oil have since been produced, for which a home use. or a n export value, has t o be found. The researches of Ehrlichr carried on since 1903 ultimately led t o the conclusion t h a t the production and composition of fusel oil depend on the character of the substance fermented, mainly with relation t o the quantities of leucine and isoleucine present. As beet molasses mash differs very widely from t h a t from potatoes or grain, it might be expected that a corresponding difference would be noted in the quantities and proportions of the amyl alcohols. The particular composition of beet molasses fusel oil is hence a question of considerable technological importance. The fusel oils of commerce differ very widely in their general characteristics, such as color. specific gravity and solubility in water, being influenced not only by the nature of the material operated on, b u t by the form of distilling apparatus, the process followed, and the care of the operator. The excise authorities generally look sharply after the presence of ethyl alcohol, and their requirements regarding this affect the presence o f other alcohols soluble in water. I n the United States, the official and commercial regulations confine the term "crude fusel oil" t o t h a t from which the lower alcohols have been so far removed by washing, t h a t a loss of not more than I O per cent. is realized by agitation with an equal bulk of water. 1

Summarized in Harden's "Alcoholic Fermentation," 1911, 75-82.

A.YD ESGIAVEERI.YG C H E X I S T R Y .

667

The term is thus used in this paper. Hence the percentage of water-soluble alcohols in fusel oil will depend t o some extent on whether the manufacturer has found it necessary to wash his oil in order t o meet the excise requirements respecting ethyl alcohol. The oil examined was amber in color. I t was found t h a t the substance or substances causing the color were soluble on shaking in dilute caustic soda. Traces of iron and copper were found, due probably to contact with these metals in the process. The specific gravity of the sample, taken by the pycnometer, a t 1 j 0 C. was 0.8370. This is somewhat higher than the figures given for fusel oil from corn, which differ widely as stated by various authorities. Worden1 gives the range for ordinary samples of washed oil as between o 810 and 0.83j. I t is evident t h a t this character will be largely influenced by the amount of water dissolved in the oil, and also by the lower alcohols which have escaped the washingout process. I n the fusel oil examined, the lower alcohols (ethyl. propyl and butyl) were roughly estimated by agitating equal volumes of oil and water thoroughly i n , a measuring tube, and noting the increase in volume of the water layer. This amounted t o 8 per cent. I n a similar manner the amount of water was determined approximately by agitating I O cc. of the oil with I O O cc. of petroleum benzine and measuring the water layer. This amounted t o j per cent. Then 5 0 0 cc. of crude unwashed fusel oil was thoroughly dried over anhydrous sodium sulphate and fractionated repeatedly until the volumes of the fractions remained constant. The result was as follows: I.

11.

-80' cc.

-/ 3 -0

i

80'-85' 18 cc.

111. 85O-9Oo 6 5 cc.

1 '.

11'. 90°-1050 1 1 cc.

105°-1150 22 cc.

VI. 115°-1200 27 cc.

1-11, VIII. IX. X. XI. 120 '-125' 125 '-128' 128 O-I3O0 130'-I33 ' residue 29 cc. 45 cc. 260 cc. 33 cc. 10 cc. TOTAL,473.5 cc. WATER, 20.0 cc. Loss, 6 . 5 cc.

A sample of the crude oil was thoroughly agitated with water, and the oil separated and dried: j o o cc. of this on fractionation gave the following: I.

11.

75'-80° none

80'-85' none

111. 85'-90' very little

VII. VIII. IX. 120°-125 O 125 O-128' 128 '-130' 4 0 cc. 75 cc. 278 cc. TOTAL, 499 cc. LOSS 1 , cc.

IV. 90°-1100 10 cc.

X. 130°-133 51 cc.

O

1'.

VI.

110°-1150 I O cc.

115°-1200 25 cc.

XI. residue 10 cc.

On comparing the results of the two fractionations i t will be noted t h a t fractions I , I1 and I11 have practically disappeared from the washed oil. These were identified, as will be shown later, as lower alcohols which are more or less soluble in water. Although the boiling points indicated the particular alcohols which were present in the various fractions it was decided t o substantiate this b y transforming the alcohols into esters. On examining the literature, i t was found t h a t the esters of acetic acid were more completely described than those of any other acid, and consequently served best for purposes of identification. Action with acetyl chloride alone was found 1

JX-orden. .Vitro-cellulose I d u s t r y . 1, 208.

668

T H E JOC'R1YA4L OF I L Y D I / S T R I A L A I Y D E.VGI,VEERISG C I I E M I S T R Y .

t o be unsatisfactory, yielding a mixture in each case difficult to fractionate, and always leaving a high boiling residue, indicating t h a t the reaction was not one of simple esterification alone. The method of Einhornl was adopted and proved very satisfactory. Accordingly I O grams of each fraction were dissolved in about 60 grams of pyridine, and a little more than the theoretical quantity of acetyl chloride was added. After standing for 6-8 hours the mixture was poured into cold dilute sulphuric acid. The ester which separated as an oil was dried over anhydrous sodium sulphate a n d fractionated. The following were the results:

Sept., 1 9 1 2

with a small quantity of active amyl ester. This fraction hencc consisted of isoamyl alcohol with some active amyl alcohol. The Residue (XI), which amounted to I O cc., was placed in a small flask and distilled. About 8 cc. came over below 150'; a very little up to 180'; and about 1 % cc. was left in the flask. This last residue appeared to have decomposed with the production of brownish vapor. The 8 cc. t h a t came over was esterified and the ester fractionated, and found for t h e greater part, to distil between 135'-140', being the boiling point of the isoamyl ester. That which was collected a t 140'-15o', and was fairly constant between 141'-142', was the active amyl ester. Very little came over above 150', b u t if a sufficient quantity had been available, would probably have been found t o consist of the hexyl and heptyl esters. This residue hence consisted of isoamyl alcohol and active amyl alcohol, with probably small quantities of higher alcohols. There were no semi-solid substances observed toward the end of the operation as has sometimes been the case with fusel oil from corn. Ten grams of fusel oil were titrated against a standard potash solution using phenolphthalein as the indicator. I O O grams of fusel oil contained 4 . 5 cc. normal acid. This is equivalent to 0 . 5 per cent. free acid in the oil, calculated on C,H,,-COOH as a basis. Ten grams of fusel oil were boiled with 75 cc. of 0 . 5 N.KOH for one hour, under a reflux condenser. After cooling, the excess of alkali was titrated against standard acid. One hundred grams of oil contained 1 8 . 4 cc. of normal free acid and acid from esters. Of this quantity, 4 . 5 cc. were free fatty acids, while the balance (13.9 cc.) represented the acids from the esters. Using CsH,lCOOCjHll, the amyl ester of capronic acid, 1 s a basis for calculation, this is equivalent to 2 . 5 per cent. ester in the oil. The bases present in the oil were separated in the usual manner. The amount was too small for purification and estimation. However, the distinctive odor of pyridine was recognized. Furfurol could not be detected.

Fraction Z (75 '-80') .-See table.-The boiling point indicated ethyl alcohol, which was confirmed by the iodoform reaction. Fraction IZ (80'-85 ').-The b. p. indicated the presence of isopropyl alcohol-b. p. 8 2 .8'. The ester on fractionation boiled chiefly a t 90°-94', the boiling point of isopropyl acetate. This fraction was hence largely isopropyl alcohol. The boiling points of fractions I11 and IV, as well as fractions V I 1 and VI11 made it difficult alone to decide with certainty as t o t h e particular alcohols present. However, upon transformation into acetic esters, and careful fractionation, i t was found possible to identify them. Fraction ZII (85 0-900).-As this fraction was very small, it was esterified and added to the ester from fraction IV. Fraction ZV (90 '-105 ').-The ester distilled chiefly between 110'-120'. I t was hence the isobutyl ester-b. p, II~-'II~'. A little isopropyl, and traces of amyl esters, were also detected. Fraction I/ (105'-115') -The b. p. indicated the presence of isobutyl alcohol-b. p. 108'. The largest portion of the ester the temperature being very conwas recovered a t IIO'-IZO', stant at 115O-117'~ indicating isobutyl ester. A very small quantity of normal butyl ester, with a boiling point of 125', was obtained. This fraction hence consisted chiefly of isobutyl Consequently, _ ' I the composition of fusel oil from alcohol with some normal butyl alcohol. beet molasses " may be summarized as follows: Fraction 111 ( I 15'-1zo').-The b . p. indicated the presence of normal butyl alcohol-b. p. I I 7 '. The main portion of the Cc. from 500 cc. Per cent. Ethyl alcohol.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1 ester came over a t I Z O ' to 130'; mostly a t 125', the boiling point Isopropyl alcohol. . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4 of t h e n. butyl ester. Some methyl normal propyl carbinol Isobutyl alcohol., . . . . . . . . . . . . . . . . . . . . . . . . . . 30 6 ester, with a boiling point of 133'-135', was recovered also. n. Butyl alcohol . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 6 This fraction hence consisted chiefly of n. butyl alcohol with Methyl + I . propyl carbinol.. . . . . 40 8 Amyl alcohols Active amyl alcohol.. . . . . . . . . 2 10 42 some methyl normal propyl carbinol. Isoamyl alcohol.. . . . . . . . . . . . . 130 26 Fraction VZZ (1~0~-125~).-l'he ester of this fraction on F a t t y acids.. . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 0 .5 distillation separated very sharply into two fractions, the larger 2.46 Esters of the fatty acids.. . . . . . . . . . . . . . . . . . . . . 12 3 with b. p. 133'-136' indicating methyl normal propyl carbinol Pyridine and other hases. . . . . . . . . . . . . . . . . . . Present I'resent Hexyl alcohols. etc. . . . . . . . . . . . . . . . . . . . . . . . Traces Traces ester, and the smaller with b. p. 138'-139' indicating isoamyl __ ester. Hence, this fraction consisted chiefly of methyl n. propyl 479.8 95.96 carbinol with some isoamyl alcohol. 20.0 4.00 U-ater. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . __ __ Fraction VIZZ (125'-128').-A small portion of ester came Total.. . . . . . . . . . . . . . . . . . . . . . . . . 499.8 99.96 over between 130' and 137', containing some methyl n. propyl carbinol ester. The largest fraction, which was very constant, This composition does not present any marked distilling between 137-140', represented the isoamyl ester. difference of practical importance from t h a t given for This fraction was chiefly isoamyl alcohol, with a little methyl corn by LeBe1,I and largely quoted, which is as foln. propyl carbinol. lows : Fraction Z X (1~8'-13o').-The b. p. indicated the presence Per cent. of active amyl alcohol, b. p. 128.5'. Practically all the ester Ic-ormal propyl alcohol.. , , . . , , , ., . 3.69 Isobutyl alcohol., . . . . . . . . . . . . . . . . . . . . . . 15.76 came over between 141' and 142', being t h e boiling point of Amyl alcohols. . . . . . . . . . . . . . . . . . . . . . . . . . . . 75.85 active amyl ester, with a small amount of the isoamyl ester, Hexyl alcohols,,. . . . . . . . . . . . . . . . . . . . . . . . . . 0 , 1 3 This fraction was hence chiefly which boils at 137'-140'. ..................... 0.16 active amyl alcohol, with a small amount of isoamyl alcohol, Fatty acid esters,, . . . . . . . . . . . . . . . . . . . . . . . 0.30 Terpines and terpine hydrates. . . . . . . . . . 0,08 Fraction X (130~-133').-The b. p. indicated the presence Purfurol. heptyl alcohol and hases . . . . . . . . . . 0 . 0 2 of isoamyl alcohol, b. p. 131.5'. Practically all the ester came __ pver between 138'-140°, proving it to be the isoamyl ester, 95.99

i

~

, ,

2

Liebig's Annulem, 301,

95 (Einhorn).

1

, ,

Maercker's "Spiritudabrication," 1,. 53.

Sept., 1912

The amount of total amyl alcohols, which, industrially, is t h e chief point, shows a diflerence of only 0.15 per cent. The respective figures for butyl alcohols differ b y 3 ~ 7 6per cent., though t h e propyl closely corresponds. This is. however, of little importance, as it shows only t h a t one sample of oil was’ washed more t h a n tbe other. TElsrorcro,

CHEMISTRE’.

T H E J O C R S A L OF I.YDc‘STRIAL A-VD EA-GISEERISG

ChXhDA.

669

added, and diluted t o jo cc. with water. No difference in color could be noticed among the three and they showed about I / ~ the color of a sample of four cc. of t h e 0.05 gram in I O O cc. sol. not treated with lead! cream. Inasmuch as lead sulphate was precipita’te’d out when the ferrous sulphate and bi-omine tvereadde3 i.t was thought possible th’at thlis pskcipitate mecKan!-’ c.ally cafried down somie vanlillid. Since vanilla contains srriall an-ioudts oi ox&. acid! and phosphates, 4 cc. of vadillin staddard ( 0 . 0 5 in; IO@) was treated with 5 drops of 85 per cent. phosphoric, acid and then wlth lead cream. I t was not necessary t o centrifuge as the lead agglutinated immediately, and on the addition of the ferrous sulphate and bromine water no precipitate was formed. The solution matched up t o a standard untreated with lead except t h a t i t was of a purplish color. But if four drops of phosphoric acid were used instead of five, a precipitate was formed on the addition of bromine water and ferrous sulphate and the color did not match an untreated standard. A series of experiments were carried on in the same manner with malic acid instead of phosphoric, and it was found t h a t 0.5 cc. of a I O per cent malic acid gave the best results ; the solution had the characteristic bluish green color instead of being purplish as with the phosphoric acid. On standing four weeks, the samples acquired a yellowish tinge in comparison to freshly prepared samples. I n the original method of MoerkI a few drops of ferrous sulphate are added and then t h e bromine drop by drop t o a maximum color. I n the method of the government, however, an excess of bromine is added and sufficient ferrous sulphate t o produce the maximum color. A number of experiments were carried out t o see if i t made any difference which way i t was done and t o find the limit in amount of ferrous sulphate and bromine water if any. One cc. of standard vanillin solution (0.05 in 100) was used in all cases. I

DIFFICULTIES IN THE COLORIMETRIC ESTIMATION OF VANILLIN. Uy \V. S. HUBBARI). Received July 19, 1912.

hloerk,? in 1891, found in testing for vanillin t h a t he had some difficulty in always obtaining the ferric chloride color reaction, and tried using a ferrous salt and oxidizing i t with bromine water. I n the latter case, he found the color much more intense than with ferric chloride and was able t o detect‘ I : I O O , O O O instead of I : 2 , 0 0 0 as before. I n a paper3 a month later he proposed a colorimetric estimation of vanillin in vanilla extracts by this means and his method with a few modifications was later adopted by the Bureau of Chemistry.4 The procedure is as follows: “Measure 2 cc. of the vanilla extract into a test tube and add about j cc. of lead hydrate; mix thoroughly, pour upon a small wet filter, collect filtrate and washings in a jo cc. graduated Kessler t u b e ; add a n excess of bromine water ( 3 or 4 drops) and sufficient freshly prepared I O per cent. ferrous sulphate solution t o produce the maximum bluish-green color t h a t will result if vanillin is present, a n d fill t o the mark with water .’ ’ “Compare with solutions containing a known amount of vanillin treated as directed above.” The lead hydrate is prepared as follows: “Dissolve 2 0 0 grams of lead acetate in 850 cc. of water, filter and add an excess of potassium hydroxide. Let the precipitate settle and wash thoroughly b y decantation with repeated portions of water until perfectly neutral. Keep in j o o cc of water in the reagent bottle, and shake t o form an emulsion-like mixture before adding t o decolorize.”,, The above method has been used in this laboratory with indifferent success for the past two years. Not until this year were the standards p u t through the treatment with lead hydrate as is directed in the method: then the troubles quickly multiplied. I t was noticed t h a t i t was not always possible t o match two standards containing the same amount of vanillin and treated in a like manner. An attempt was made t o locate the trouble by means of a number of experiments. Four cc. each of three solutions of vanillin standard made up according t o t h e directions given in Bulletin 1 0 7 , but containing o . o j , 0 . 1 0 and 0 . 2 0 gram o f vanillin per I O O cc. respectively, were treated with lead cream, centrifuged and filtered, ferrous sulphate and bromine Read a t the February meeting of University of Michigan Section of the:.%. C. S.y I wish to thank Dean J ,0 . Schlotterbeck for his interest in my work. Am. J . P h a ~ v r .63, , 521. i Ibid.. 572. Hir//. 107, (r

.i 2

5

5 4

5 5

5 6

5 7

5 8

Drops Br sol. Drnps FeSO, sol.

1

3

,

h-o color

Darkest and equal

._-_--

Lighter and e r i u a l

I t was found t h a t after obtaining the maximum color in the above tests and then diluting t o jo cc. with water, it was necessary in every case t o add at least 2 drops more bromine water t o bring the color up to a maximum, a n d this could not be accomplished by adding t h e extra 2

A m J . P h a r n , , 63, 5 2 1 .